US scientists believe they could use brain stem cells to cure diabetes.
Transplants could ultimately remove the need for daily insulin injections
Although the work is not yet ready to be tested on human patients, results in animals have been promising, say the Stanford University researchers.
They were able to coax the immature brain cells to develop into the insulin-producing islet cells that are lacking in diabetes.
Eventually, these could be used for curative transplants, the scientists told the journal PLoS Medicine.
Scientists have already been looking at using stem cells taken from embryos to treat diabetes.
These are primitive "master" cells that can be programmed to become many kinds of tissue.
However, there have been concerns that these cells can turn cancerous, are difficult to work with in the laboratory.
Dr Seung Kim and colleagues looked at whether stem cells taken from the brain might work just as well and avoid some of these issues.
Dr Kim said: "When you look at islet cells you realise that they resemble neurons."
In some insects, such as fruit flies, the cells that produce insulin and regulate blood sugar are also neurons.
Dr Kim's team found that when they added a cocktail of chemicals to brain stem cells, taken from aborted fetuses, the cells changed and, although they were not identical to islet cells, they were able to produce insulin in response to blood sugar levels.
To find out whether these cells would work, they transplanted them into a cavity in the kidney in mice where other types of insulin-producing cells have been found to survive before.
When the blood sugar went up in these mice, the transplanted "mature" brain stem cells again released insulin.
Four weeks later, the cells were still alive and producing insulin and none had turned cancerous.
Dr Kim said, although it was early days, the work suggested that stem cells could be used to replace islet cells and free people with type 1 diabetes daily insulin injections.
Some patients have already received transplants using islet cells taken from living relatives or dead donors.
Dr Angela Wilson, director of research at Diabetes UK said: "This is an interesting result and may provide another avenue to explore in our search for a cure for diabetes.
"However, the work is in the very early stages of development and has yet to be reproduced in humans.
"We'll certainly be following the progress of this research with interest."